JP7759367B2 - Program and system - Google Patents

Description

The present invention relates to a program and a system.

Technologies for supporting the creation of computer-based games (hereinafter simply referred to as "games") have been known in the past.

For example, images representing objects such as the main character, enemies, or items in a game are generated, and the images of multiple objects are composited together. The composite image for the display screen is then displayed in synchronization with the display cycle. The speed at which the object acts is also specified. In this way, techniques are known that add novelty to existing image processing (see, for example, Patent Document 1).

Japanese Patent Application Laid-Open No. 2003-340145

However, with conventional technology, events that occur in a game are predetermined by a program or the like. That is, for example, when playing the same game repeatedly (so-called "loop play"), the user ends up clearing the same events. As a result, users who play repeatedly tend to get bored, which poses a problem of losing interest in the game.

The purpose of this invention is to prevent users from becoming bored with the game, even when playing it repeatedly, such as in loop play, and to prevent the game's enjoyment from declining.

In order to solve the above problems according to the present invention, a program
Computer,
a first event data input means for inputting first event data indicating a first event to be generated in the game;
an unknown data input means for inputting unknown data indicating a play whose correct answer is unknown to a trained model trained using training data including the first event data and log data;
an execution means for generating second event data indicating a second event to be caused to occur in the game separately from the first event based on the unknown data, and for generating output data indicating a story of the game using the first event data and the second event data;
Based on the output data, the user terminal used by the user who plays the game is made to function as a playing means that allows the user to play the game.

According to the present invention, even when a game is played repeatedly, such as in a loop playthrough, it is possible to prevent the user from becoming bored with the game and prevent the game's enjoyment from declining.

FIG. 1 is a diagram illustrating an example of a system configuration according to an embodiment of the present invention. FIG. 10 is a diagram illustrating an example of pre-processing. FIG. 10 is a diagram illustrating an example of execution processing. FIG. 1 is a diagram illustrating an example of the overall process of AI learning and execution. FIG. 2 is a diagram illustrating an example of a hardware configuration of an information processing device. FIG. 1 is a network diagram showing an example of an AI configuration. FIG. 10 is a diagram illustrating an example of a game configuration. FIG. 10 is a diagram illustrating an example of a learning process in a novel game. FIG. 10 is a diagram illustrating an example of execution processing in a novel game. FIG. 10 is a diagram showing an example of output from a novel game. FIG. 10 is a diagram illustrating an example of overall processing. FIG. 2 is a diagram illustrating an example of a functional configuration. FIG. 10 is a diagram illustrating a configuration example using an auxiliary device.

The following describes the embodiment with reference to the drawings.

[System configuration example]
1 is a diagram showing an example of a system configuration according to this embodiment. For example, as shown in FIG. 1, a system 1 mainly includes user terminals 20A, 20B, and 20C (hereinafter, these may be collectively referred to as "user terminals 20") and a server 11.

Hereinafter, the person who manages server 11 will be referred to as "administrator 5." Furthermore, the people who operate user terminals 20A, 20B, and 20C will be referred to as "user 4A," "user 4B," and "user 4C" (hereinafter, these will sometimes be collectively referred to as "user 4").

Administrator 5 is responsible for managing the information processing service provided by System 1. On the other hand, User 4 is a user of the information processing service provided by System 1. Furthermore, Administrator 5 and User 4 differ in which information processing device they operate: Server 11, an example of a management device, or User terminal 20. Hereinafter, User 4 will be the player of the game, and Administrator 5 will manage the game and Server 11.

Note that the example shown in Figure 1 has three user terminals 20 and one server 11, but the number of servers 11, the number of user terminals 20, the number of administrators 5, and the number of users 4 are not important.

The server 11 and the user terminal 20 are connected to each other so that they can communicate with each other via a communication network 2. For example, the communication network 2 is the Internet, a mobile communication system (such as a public line using 4G (4th Generation, fourth generation mobile communication standard) or 5G (5th Generation, fifth generation mobile communication standard)), a wireless network such as Wi-Fi (registered trademark), or a combination of these.

The user terminal 20 downloads a program for playing the game (hereinafter referred to as the "game program") from the server 11, or accesses the server 11 to provide a game service. Note that communication with the server 11 is not required to play the game. In other words, the user terminal 20 may download the program or install it from media to create an environment for playing the game.

[Examples of AI (Artificial Intelligence) Learning and Execution]
Hereinafter, AI learns through "pre-processing." The AI in the learning stage, i.e., "pre-processing," is referred to as the "learning model A1." Then, once learning has progressed to a certain extent, the learning model A1 becomes the "trained model A2." Hereinafter, the execution stage in which output processing is performed using the trained model A2 is referred to as the "execution process."

"Pre-processing" is performed before "execution processing." However, "pre-processing," i.e., trained model A2 may continue to train, before "execution processing" is performed.

[Example of pre-processing]
2 is a diagram showing an example of pre-processing. For example, the pre-processing is performed by the server 11.

Learning model A1 inputs learning data D1 and performs learning. In other words, learning model A1 performs so-called "supervised" learning.

Learning data D1 includes known log data and is data that corresponds to the "correct answer" for this known log data. Specifically, learning data D1 includes log data that indicates the game play results and play process and for which the "correct answer" is known (hereinafter simply referred to as "log data D11"), as well as correct answer data D12. First event data DE1 is also input.

The learning model A1 learns the correspondence between inputs such as log data D11 and outputs indicated by the correct answer data D12 based on the input of learning data D1.

Details of the log data D11, first event data DE1, and correct answer data D12 will be described later.

Furthermore, it is desirable for the learning model A1 to be trained using big data D4. For example, big data D4 is data on the Internet. However, big data D4 may also be data entered by an administrator 5 or the like. In this way, by training using big data D4, it is possible to create AI that generates natural sentences that sound like they are written by a human.

[Example of execution process]
3 is a diagram showing an example of the execution process. For example, the execution process is performed by a plurality of information processing devices, such as the user terminal 20 and the server 11, or the user terminal 20 and the server 11, working together.

Trained model A2 is the state in which training model A1 has been trained through pre-processing. In other words, when the pre-processing shown in Figure 2 is performed, trained model A2 is generated.

When unknown data D2 is input, trained model A2 generates output data D3 for the unknown data D2.

Unknown data D2 is log data for which the "correct answer" is unknown (hereinafter referred to as "unknown log data D21").

Once the output data D3 is generated, the output data D3 is transmitted to, for example, the user terminal 20. The user terminal 20 then outputs an output screen or the like to the user 4 based on the output data D3.

Details of the unknown data D2, unknown log data D21, output data D3, and output based on the output data D3 will be described later.

Figure 4 shows an example of the overall process for AI learning and execution. The relationship between the pre-processing shown in Figure 2 and the execution process shown in Figure 3 is as shown in Figure 4.

Note that the pre-processing and execution processing do not have to be performed in the sequential order illustrated in the figure. Therefore, it is not necessary to have the period during which preparations are made through pre-processing and the subsequent period during which execution processing is performed consecutively. Therefore, once the trained model A2 has been created, the execution processing may be performed after some time has passed since the pre-processing. Furthermore, once the trained model A2 has been generated, the execution processing may be performed by reusing the trained model A2.

The learning data D1 and unknown data D2 are different between the learning process and the execution process. Furthermore, during the learning stage, the AI is a learning model A1, but once learning has progressed to a certain extent, it becomes a trained model A2. In this way, the trained model A2, which has been trained using big data D4 and the like as training data, is what is known as "generative AI."

In the learning data D1, the "correct answer" is known, whereas in the unknown data D2, the "correct answer" is unknown. Specifically, since the learning data D1 includes the correct answer data D12, the correct answer data D12 is associated with the log data D11, whereas the unknown data D2 does not include the correct answer data D12. Therefore, in the learning process, the relationship between the log data D11 and the correct answer data D12 is known. Therefore, when the learning data D1 is input, the learning model A1 can learn the correlation between the log data D11 and the "correct answer."

On the other hand, the unknown data D2 does not contain the correct answer data D12, and the "correct answer" for the unknown data D2 is unknown. The trained model A2 then generates output data D3 for the unknown data D2 based on the correlation between the training data D1 learned in pre-processing and the correct answer data D12.

The execution process may partially use tables, etc. In this way, in a configuration that uses tables, or is so-called rule-based, the pre-processing is a process that prepares for inputting a table (also called a lookup table (LUT)) or a mathematical formula, etc.

[Example of hardware configuration of information processing device]
5 is a hardware configuration diagram of an information processing device. The information processing device is a server 11, a user terminal 20, etc. Hereinafter, it is assumed that the information processing device has the same hardware configuration as the server 11. For example, the information processing device is a workstation or a general-purpose computer such as a personal computer. However, each information processing device may have a different hardware configuration.

The server 11 mainly comprises a processor 111, memory 112, storage 113, an input/output interface 114, and a communication interface 115. Furthermore, each component of the server 11 is connected to a communication bus 116.

The processor 111 performs processing and control by executing a series of instructions contained in the server program 11P stored in the memory 112 or storage 113.

The processor 111 is an arithmetic device and control device, such as a CPU (Central Processing Unit), GPU (Graphics Processing Unit), MPU (Micro Processing Unit), FPGA (Field-Programmable Gate Array), ASIC (Application Specific Integrated Circuit), or a combination of these.

Memory 112 is a main storage device that stores server program 11P, data, etc. For example, server program 11P is loaded from storage 113. Data includes data input to server 11 and data generated by processor 111. For example, memory 112 is RAM (Random Access Memory) or other volatile memory.

Storage 113 is an auxiliary storage device that stores the server program 11P and data, etc. Storage 113 is, for example, a ROM (Read-Only Memory), a hard disk drive, flash memory, or other non-volatile storage device. Storage 113 may also be a removable storage device such as a memory card. As another example, storage 113 may be an external storage device. With this configuration, for example, in situations where multiple user terminals 20 are used, such as in an amusement facility, it becomes possible to perform updates to the server program 11P or data collectively.

The input/output interface 114 is an interface that connects external devices such as a monitor, input device (e.g., a keyboard or pointing device), external storage device, speaker, camera, microphone, and sensor to the server 11.

The processor 111 also communicates with external devices through the input/output interface 114. The input/output interface 114 may be, for example, a Universal Serial Bus (USB), a Digital Visual Interface (DVI), a High-Definition Multimedia Interface (HDMI (registered trademark)), wireless, or other terminal.

The communication interface 115 communicates with other devices (e.g., user terminal 20) connected to the communication network 2. For example, the communication interface 115 is a wired communication interface such as a LAN (Local Area Network), or a wireless communication interface such as Wi-Fi (Wireless Fidelity), Bluetooth (registered trademark), or NFC (Near Field Communication).

However, the information processing device is not limited to the above hardware configuration. For example, the user terminal 20 may further include a sensor such as a camera. Various data acquired by the user terminal 20 via the sensor may then be transmitted to the server 11.

[Examples of the configuration of a training model and a trained model]
6 is a network diagram showing an example of an AI configuration. The learning model A1 and the trained model A2 are, for example, AIs having a configuration shown in the following network.

In the following, we will explain an example in which the learning model A1 and the trained model A2 are implemented on the server 11, i.e., on the cloud. However, some or all of the learning model A1 and the trained model A2 may also be implemented on the user terminal 20, etc.

The network 300 has, for example, an input layer L1, an intermediate layer L2 (also called a "hidden layer"), and an output layer L3.

The input layer L1 is the layer that inputs data.

The intermediate layer L2 converts the data input by the input layer L1 based on weights, biases, etc. The results of this processing in the intermediate layer L2 are transmitted to the output layer L3.

The output layer L3 is the layer that outputs the output content, etc.

Then, through learning, the weight coefficients and parameters changed through learning are optimized. Note that network 300 is not limited to the network structure shown in the figure. In other words, AI may be realized through other machine learning methods.

For example, AI may be configured to perform preprocessing such as dimensionality reduction (for example, a process that converts relationships of three or more dimensions into relationships that can be determined with simple calculations of three or fewer dimensions) using "unsupervised" machine learning. It is desirable for the relationship between input and output to be processed using simple calculations such as linear expressions. Calculations of this nature can reduce computational costs.

The AI may also undergo processes such as dropout to reduce overfitting (also known as "overfitting" or "over-fitting"). Other preprocessing steps such as dimensionality reduction and normalization may also be performed.

AI may have a network structure such as a CNN (Convolutional Neural Network). Other network structures, such as a LLM (Large Language Model), an RNN (Recurrent Neural Network), or an LSTM (Long Short-Term Memory), may also be used. In other words, AI may have a network structure other than deep learning.

Also, the AI may be configured with hyperparameters. In other words, the AI may be configured so that some of its settings are performed by a user, etc. Furthermore, the AI may specify the features to be learned, or the user may set some or all of the features to be learned.

Furthermore, learning model A1 and trained model A2 may utilize other machine learning methods. For example, learning model A1 and trained model A2 may be unsupervised models that undergo preprocessing such as normalization. Furthermore, learning may be reinforcement learning (a learning method in which an AI makes a choice and is given an evaluation (reward) for the choice, resulting in a larger evaluation), etc.

During learning, data expansion may be performed. That is, in order to increase the amount of learning data used in learning learning model A1, preprocessing may be performed to expand a single piece of experimental data, etc., into multiple pieces of learning data. Increasing the amount of learning data in this way allows for further learning of learning model A1.

Furthermore, the learning model A1 and the trained model A2 may be configured to perform transfer learning or fine tuning. In other words, since each user terminal 20 often has a different execution environment, the settings for each device may differ depending on the execution environment. For example, the basic configuration of the AI may be trained on a separate information processing device. Thereafter, each information processing device may undergo additional training or configuration to optimize it for its respective execution environment.

[Example of Overall Game Configuration, Example of First Event, and Example of Second Event]
7 is a diagram showing an example of the structure of a game. For example, a game may be composed of multiple events. Note that a game may also include elements other than events. However, for the sake of simplicity, the following description will be given using an example in which the game is composed only of events.

Events are objects that players must overcome within the game, such as troubles, opportunities, incidents, occurrences, or occasions that arise as the game progresses.

Hereinafter, there will be two types of events: "First Event E1" and "Second Event E2." Therefore, the game will consist of a combination of First Event E1 and Second Event E2.

The content of the first event E1 is set by the first event data DE1. For example, the first event data DE1 is set in advance by the administrator 5 or the like. In other words, the first event E1 is an event for which a so-called "plot" (also called a "scenario plot") is prepared in advance. Furthermore, the first event E1 occurs in every play (when the same user 4 plays repeatedly, or when different users 4 play, etc.), and the content is the same. Furthermore, the first event E1 is a fixed event that is always set so that the user 4 will play it.

In this way, the first event E1 is a "mandatory event" that must be overcome in every playthrough. Therefore, a mandatory event is, for example, an event that includes an important scene, the appearance of an important character, or the acquisition of a required item.

Note that the first event E1 does not have to be an event that is completely set in advance. For example, depending on the content of the second event E2 that is input before and after the first event E1, the first event E1 may be changed to ensure consistency between the connecting parts of the events.

Hereinafter, the game will be described as having multiple first events E1, such as a first required event E11, a second required event E12, etc. The conditions, timing, and order of occurrence of each first event E1, such as the first required event E11, the second required event E12, etc., are set by the first event data DE1.

The content of the second event E2 varies depending on the log data D11. For example, the second event E2 is an event that occurs between the first required event E11 and the second required event E12.

However, the second event E2 is not limited to occurring between the first required event E11 and the second required event E12, but may also be an event occurring within the first required event E11 (sometimes called a "sub-event" or "minor event").

In this way, the second event E2 is a variable event whose content is set by the log data and can be changed by user 4.

For example, the second event E2 is a "variable event" that changes depending on the results of play up to that point (for example, the results of selecting options in the game), the characters that appear in the game, or the profile of the user 4. Furthermore, not only the content of the second event E2 but also the timing of its occurrence may be variable. Alternatively, the second event E2 may be an event that occurs before the first required event E11. However, the timing of the occurrence of the second event E2, its occurrence probability, etc. may be set (including partially set) in advance by the administrator 5, etc.

Note that the second event E2 may be an "optional" event where user 4 can choose whether or not to participate in the event, or an event where completion is required. Furthermore, two or more second events E2 may be generated.

[Example of learning process in a novel game]
8 is a diagram showing an example of learning processing in a novel game. The following describes an example of a novel game.

As the player progresses through a novel game, text is first output to the screen. Next, User 4 performs an operation to select one option from multiple options (hereinafter referred to as a "selection operation"). In this way, a novel game is a game in which the story that follows the game branches depending on the selection operation, and the ending of the story changes depending on the selection operation. For example, a novel game is a game in which the player progresses with the objective of identifying the culprit in a mystery novel story.

[Example of log data]
The log data D11 is generated as the game progresses as the game is played. That is, the log data D11 is data that indicates the game play process. The log data D11 may be generated as the game progresses, or when a save or the like is performed, data from the previous save point to the most recent save point may be generated all at once.

The log data D11 is a data structure that includes, for example, play data D111, image data D112, and profile data D113. However, each piece of data does not have to be in separate files, and a single piece of data may contain multiple pieces of content.

The play data D111 is data that indicates the content of the gameplay. For example, the play data D111 is the content of a selection operation, i.e., the option selected from multiple options. The play data D111 may also include related information such as items obtained, parameter changes, events that occurred, the content of the events, play results, performance, time spent, other participants, or characters that appeared. The play data D111 may also include parameters that are not visible to the user 4.

Image data D112 is data such as a recorded image (the image is a still image or a video) of an output screen showing play by user 4 (the output screen may be from the perspective of user 4 or another person). Image analysis processing may also be performed on image data D112, and various data may be extracted and used as auxiliary data. Specifically, auxiliary data is what is known as a tag. For example, as image analysis processing, objects shown in image data D112 are recognized, and the name of the object is output as a tag. Note that image analysis processing is not limited to object recognition, and other types of analysis may also be performed.

In addition, the auxiliary data may be associated with log data D11, etc. In other words, the log data D11 does not have to be a single type of data, but may be a data configuration that includes associated auxiliary data, etc.

For example, the subject of image data D112, particularly in video, varies from frame to frame. Therefore, the contents of the corresponding log data D11 and the auxiliary data containing the results of the image analysis process also change. Furthermore, the correspondence between the generated auxiliary data and the multiple pieces of log data D11 and multiple pieces of image data D112 can become more complex as the number of pieces of data increases. Therefore, auxiliary data can be easily managed if it is associated with an associated piece of data that indicates which piece of data it corresponds to. For example, the association can be achieved by including the name of the data that was the subject of analysis in the auxiliary data.

Profile data D113 indicates the profile of a character appearing in the game or of user 4. Which character or user 4's profile the profile data D113 indicates varies depending on the game settings, input items, etc.

For example, the profile data D113 is entered by the user 4 when the game starts. However, the profile data D113 may not be set by the user 4 but may be set in advance, or may be obtained from the user terminal 20 or another database, etc.

Specifically, profile data D113 includes setting items such as name, birthplace, gender, age, occupation, and friendships. The profile items are set in advance.

The name may be, for example, the real name, abbreviation, or pseudonym of user 4. The name may also be identification information such as a nickname set for the character.

Similarly, the origin, gender, age, occupation, and friendships may be user 4's settings or fictional settings in the game.

When the log data D11 described above is input and is also input as correct answer data D12 containing the second event E2, which is the correct answer, the learning model A1 can learn the correlation between the log data D11 and the correct answer, i.e., the second event E2.

By learning in this way, a trained model A2 can be generated that generates a second event E2 based on the log data D11. Furthermore, when the first event data DE1 is input, the first required event E11 and the second required event E12 can be set to occur in the game. Therefore, a trained model A2 that generates a game that combines the first required event E11, the second required event E12, and the second event E2 is generated through learning.

[Example of execution process in a novel game]
9 is a diagram illustrating an example of an execution process in a novel game. After a trained model A2 is generated by the learning process shown in FIG. 8, the execution process is performed as follows.

Unknown log data D21 is input during the execution process. Compared to the learning process, the execution process differs in that unknown log data D21 is input and there is no correct answer data D12. In addition, the trained model A2 outputs output data D3, etc.

When the unknown log data D21 is input, the correct answer is unknown. Then, in the execution process, the trained model A2 generates second event data DE2 based on the correlation between the trained log data D11 and the correct answer data D12, and taking into account the unknown log data D21. Then, the trained model A2 uses the second event data DE2 to create a game that serves as a story in which the second event E2 occurs.

[Example of output from a novel game]
10 is a diagram showing an example of output from a novel game. For example, the game progresses in the order of a first required event E11, a second event E2, and a second required event E12. Note that the profile is assumed to have been entered in advance at the start of the game.

For example, in the first required event E11, a selection operation is performed to select one option from options 30. Specifically, options 30 are "option A" and "option B," and the lines or actions of the character in the game are determined by the operation of user 4. The result of the selection operation is then input to trained model A2 as unknown log data D21.

The second event E2 is composed of, for example, a first sentence 31, a second sentence 32, and a third sentence 33. The first sentence 31, the second sentence 32, and the third sentence 33 are generated so as to change depending on the unknown log data D21, i.e., the result of the selection operation, the profile, etc.

Specifically, first sentence 31 is a sentence that reflects the "Origin: Country XX" setting in the profile. Furthermore, as first sentence 31 indicates, a setting is generated in which the "Origin: Country XX" of a character appearing, "Character XXX," is "Origin: Country XX." In particular, if the GUI (Graphical User Interface) used to enter "Origin" in the profile is in text format rather than a selectable format such as a pull-down menu, it is difficult to prepare a sentence like first sentence 31 in advance. Therefore, first sentence 31, which uses the "Origin: Country XX" setting in the profile, is used in second event E2.

The second sentence 32 is a sentence that reflects the result of the selection operation for option 30 in the first required event E11. Specifically, if a selection operation to select "option B" is not performed in the first required event E11, the second sentence 32 will not be output.

Third sentence 33 is an example of a sentence that indicates a "clue" that is advantageous for clearing the game. For example, the "clue" indicated by third sentence 33 can only be obtained if "Origin: Country X" is set in the profile and "Option B" is selected in the first required event E11.

The second event E2 is an event in which a character named "Character◆◆" appears, who only appears in events where game hints can be obtained, such as the third sentence 33. In other words, the character "Character◆◆" does not appear in other stories, and is not a character prepared in advance by a program or the like.

A character called "Character◆◆" is generated for each story. Specifically, even in the same scene (for example, the same scene in the first and second playthroughs of a loop), a character called "Character◆◆" will have different names, lines spoken, personalities, or the timing of their appearance. Furthermore, there may be stories in which a character called "Character◆◆" does not appear. In this way, a character called "Character◆◆" is a character that is created and appears by the AI for each story.

Furthermore, the first sentence 31 and the second sentence 32, which are conversations between the "protagonist XXX" controlled by user 4, do not appear in other stories, and are not sentences prepared in advance by a program or the like.

The first sentence 31 and the second sentence 32 are generated for each story. Specifically, even in the same scene, the lines spoken by the characters "Main Character XXX" and "Character ◆◆" in the first sentence 31 and the second sentence 32 will vary. For example, even if the lines have the same content, the tone or ending of the lines will vary depending on the flow of the story, gender, age, etc.

In this way, the first sentence 31, the second sentence 32, and the third sentence 33 are not generated by applying a profile or the like to a pre-prepared sentence (hereinafter referred to as a "format"), but are generated from scratch, without a format. Furthermore, the first sentence 31, the second sentence 32, and the third sentence 33 are generated so as to be consistent with the sentences before and after them.

As described above, completing the second event E2 will result in a reward, such as obtaining information that will be useful in clearing the game. Note that the result of the second event E2 is not limited to obtaining information, and other rewards may also be available. For example, the result of the second event E2 may be a reward such as obtaining an item or having a specific character join your team.

The second event E2 may also include options. That is, the second event E2 may be generated so that the ending varies depending on the selection of options within the event. For example, the second event E2 may not end in a way that provides valuable information upon clearing the game, as shown in FIG. 10, but may end in a way that does not provide valuable information even if the second event E2 is cleared, depending on the options selected.

After the second event E2 is completed, the second required event E12 occurs. For example, the second required event E12 is an event that can be completed without going through the second event E2.

It is therefore desirable that the second required event E12 be cleared no matter what second event E2 is completed. In other words, because the content of the second event E2 varies, depending on the play, there may or may not be specific information or items when the second required event E12 is reached. For this reason, it is desirable that the second required event E12 be configured so that it can be completed, at the very least, with the content of the first required event E11.

[Overall processing example]
11 is a diagram showing an example of overall processing. In the following example, the overall processing is performed successively as a pre-processing and an execution processing. Specifically, the pre-processing is steps S01 to S03. The execution processing is steps S04 to S07. However, the overall processing may include other processing steps.

In step S01, the server 11 inputs the first event data DE1.

In step S02, the server 11 inputs the learning data D1.

In step S03, the server 11 trains the learning model A1 using the first event data DE1 and the learning data D1 to generate a trained model A2.

Steps S01 to S03 constitute the learning process shown in Figure 8. As described above, after a certain amount of learning processing has been performed, learning model A1 is trained to become trained model A2. Next, execution processing is performed as follows using trained model A2 generated by the learning processing.

In step S04, the server 11 inputs the first event data DE1. For example, step S04 is performed in the same manner as step S01. Note that if the content of the first event data DE1 is the same as during the learning process, the first event data DE1 input in step S01 may be reused and step S04 may be omitted.

On the other hand, if the content of the first event E1 differs between the learning process and the execution process, new first event data DE1 with content different from that in step S01 is input in step S04.

In step S05, the server 11 inputs unknown data D2.

In step S06, the server 11 generates output data D3.

In step S07, the server 11 causes user 4 to play the game on the user terminal 20 based on the output data D3.

Steps S04 to S07 constitute the execution process shown in Figure 9. As described above, when the execution process is performed, the trained model A2 outputs output data D3 based on the unknown data D2. Using this output data D3, for example, a game (as a whole) such as that shown in Figure 7 is constructed, and when the game is distributed to the user terminal 20, user 4 can play the game.

In the above example of the overall processing, the server 11 is primarily responsible for the processing, but the overall processing may be configured so that all or part of it is executed by another information processing device, such as the user terminal 20. Therefore, the overall processing may be executed by an information processing device other than the server 11.

[Example of functional configuration]
12 is a diagram illustrating an example of a functional configuration. For example, the system 1 is an AI game event generation system including a learning device 61 and an execution device 62.

The learning device 61 includes a first event data input means 1F11, a learning data input means 1F12, and a learning means 1F13.

The first event data input means 1F11 performs a first event data input procedure to input first event data DE1 indicating the first event E1. For example, the first event data input means 1F11 is realized by the input/output interface 114, etc.

The learning data input means 1F12 performs a learning data input procedure for inputting learning data D1 including log data D11 and correct answer data D12. For example, the learning data input means 1F12 is realized by a communication interface 115, etc.

The learning means 1F13 performs a learning procedure in which the learning model A1 is trained using the learning data D1 to generate a trained model A2. For example, the learning means 1F13 is realized by the processor 111, etc.

The execution device 62 includes a first event data input means 1F11, an unknown data input means 1F21, an execution means 1F22, and a play means 1F23.

The first event data input means 1F11 in the execution device 62 is, for example, similar to the first event data input means 1F11 in the learning device 61.

The unknown data input means 1F21 performs an unknown data input procedure to input unknown data D2, including unknown log data D21, to the trained model A2 generated by the training means 1F13. For example, the unknown data input means 1F21 is realized by a communication interface 115, etc.

When unknown data D2 is input, the execution means 1F22 generates second event data DE2 based on the unknown data D2. The execution means 1F22 then performs an execution procedure to generate output data D3 that indicates the game story using the first event data DE1 and the second event data DE2. For example, the execution means 1F22 is realized by the processor 111, etc.

The play means 1F23 performs a play procedure that allows the user 4 to play the game on the user terminal 20 based on the output data D3. For example, the play means 1F23 is realized by the processor 111, etc.

The learning device 61 and the execution device 62 are, for example, servers 11. In this way, the learning device 61 and the execution device 62 may be combined into a single information processing device. Of course, the learning device 61 and the execution device 62 may also be different information processing devices.

With the above configuration, it is possible to generate a different second event E2 for each play in response to the first event E1, which is a required event. Furthermore, since the second event E2 is an event that reflects the personality of user 4 based on their profile, etc., it can be an event unique to user 4. Furthermore, if the second event E2 varies for each play, even for the same player, it is possible to increase the replayability of the game.

For example, if a method using formats or the like is used to generate the lines spoken by characters, the only change possible is to fit the name into the format, and since the basic story remains the same when the game is played repeatedly, User 4 will end up playing almost the same story and will easily become bored with the game. In particular, when formats are used, only one story is prepared for each anticipated pattern, so User 4 will often only be able to play a smaller number of stories than if second event E2 were generated by AI.

On the other hand, if the second event E2 is an event that changes with each play, User 4 can play an event whose content changes with each play. Therefore, even if User 4 plays the game repeatedly, User 4 will not get bored of the game. In this way, an attractive game can be provided to User 4.

[Modification]
The type of game is not limited to a novel game, and may be, for example, a role-playing game (RPG), a sports game, an action game, a puzzle game, or a fighting game.

For example, even in an RPG, important events are set as required events, even when playing repeatedly. Therefore, if a first event E1 is set as a required event, the user 4 will always play the first event E1. On the other hand, as the game progresses, a second event E2 will occur at any timing. For example, the items that can be obtained in the second event E2, the characters that appear, the character's lines, or the enemies that can be fought will differ for each play. Therefore, it is possible to provide an RPG in which the types of events that occur will differ for each play.

For example, in a fighting game, the second event E2 may involve subsequent opponents or characters who will become subsequent opponents (e.g., characters with different appearances or parameters) varying depending on the player's profile, score, etc. On the other hand, in the first event E1, the player may fight the same fixed opponent even when playing repeatedly.

[Other embodiments]
In the above example, the information processing device performs both pre-processing for the learning model and execution processing using the learned model. However, the pre-processing and execution processing do not have to be performed by the same information processing device. Furthermore, the pre-processing and execution processing do not have to be consistently performed by a single information processing device. In other words, each process and data storage may be performed by an information system or the like configured with multiple information processing devices.

Note that the learning process may be performed after the execution process or as an additional process before the execution process.

The above processing may be performed auxiliary by an information processing device other than the server 11 and the user terminal 20.

Figure 13 shows an example configuration using an auxiliary device. Compared to the example shown in Figure 1, the configuration shown in Figure 13 differs in that an auxiliary device 60 is added. Note that the auxiliary device 60 may be used temporarily.

The auxiliary device 60 is an information processing device installed near the user terminal 20 (in this example, it is installed near the user terminal 20A, but it may be installed near another device). The auxiliary device 60 then executes certain processes in part or in whole on behalf of the user terminal 20 or the server 11.

For example, the auxiliary device 60 may be equipped with a device specialized for graphics processing and perform graphics processing at high speed. In this way, so-called edge computing may be performed by installing the auxiliary device 60 or the like. In this way, the above-described processing may be performed by utilizing the hardware resources of various information processing devices. Therefore, the above-described processing may be performed by an information processing device different from the one described above.

The above-described processes and the data used in the processes performed in this embodiment may be executed and stored by an information processing system. For example, the information processing system may execute or store data on multiple information processing devices to achieve redundant, distributed, parallel, or a combination of these processes or storage. Therefore, the present invention may be realized in devices other than those with the hardware configurations described above and in systems other than those described above.

Furthermore, the program according to the present invention is not limited to a single program, but may be a collection of multiple programs. Furthermore, the program according to the present invention is not limited to being executed by a single device, but may be shared and executed by multiple information processing devices. Furthermore, the division of roles among the information processing devices is not limited to the above-mentioned example. In other words, some or all of the above-mentioned processing may be executed by an information processing device different from the above-mentioned information processing device.

Furthermore, some or all of the means implemented by the program may be implemented in hardware such as an integrated circuit. Furthermore, the program may be provided recorded on a non-transitory recording medium that is readable by a computer. Examples of recording media include hard disks, SD cards (registered trademark), optical discs such as DVDs, and servers on the Internet. Therefore, the program may be distributed via telecommunications lines such as the Internet.

In addition, the information processing devices that make up the information processing system may be located overseas. In other words, the information processing devices that perform some of the processing performed by the information processing system may be located overseas.

The present invention is not limited to the embodiments exemplified above. Therefore, the present invention allows for the addition or modification of components within the scope of the technical gist. Therefore, all technical matters included in the technical concept described in the claims are within the scope of the present invention. The embodiments exemplified above are preferred specific examples for implementation. Furthermore, those skilled in the art will be able to realize various modifications from the disclosed content, and such modifications are within the technical scope described in the claims.

1: System A1: Learning model A2: Learned model 1F11: First event data input means 1F12: Learning data input means 1F13: Learning means 1F21: Unknown data input means 1F22: Execution means 1F23: Play means 4: User 5: Administrator 11: Server 20: User terminal 30: Option 31: First sentence 32: Second sentence 33: Third sentence 60: Auxiliary device 61: Learning device 62: Execution device D1: Learning data D11: Log data D111: Play data D112: Image data D113: Profile data D12: Correct answer data D2: Unknown data D21: Unknown log data D3: Output data D4: Big data DE1: First event data DE2: Second event data E1: First event E11 : First required event E12 : Second required event E2 : Second event

Claims (3)

A computer including a learning device and an execution device,
The learning device
a first event data input means for inputting first event data indicating a first event to be generated in the game;
functioning as a learning means for generating a trained model that learns a correlation between the log data and the second event using training data including the first event data, log data, and correct answer data indicating a second event for the log data;
The execution device
a first event data input means for inputting the first event data;
an unknown data input means for inputting unknown data, which is unknown log data, to the trained model;
an execution means for generating second event data indicating the second event based on the correlation and taking the unknown data into consideration when the unknown data is input and output data is output, and for generating the output data indicating the story of the game using the first event data and the second event data;
causing a user terminal used by a user playing the game to function as a playing means for allowing the user to play the game based on the output data;
The first event is
a fixed event whose content is set by the first event data and which the user will definitely play,
The second event is
The content of the event is set based on the log data, and the event is variable and may vary depending on the user, and is generated differently for each play.
The second event is an event that occurs between a plurality of the first events.
program. The log data is
2. The program according to claim 1, further comprising play data indicating the content of play in the game, image data of an output screen in the game, or profile data indicating a character in the game or a profile of the user. A system including a learning device and an execution device,
The learning device
a first event data input means for inputting first event data indicating a first event to be generated in the game;
learning data input means for inputting learning data including the first event data, log data indicating play of the game, and correct answer data indicating a second event for the log data;
a learning means for using the learning data to train a learning model that learns a correlation between the log data and the second event to generate a trained model;
The execution device
a first event data input means for inputting the first event data;
an unknown data input means for inputting unknown data, which is unknown log data, to the trained model;
an execution means for generating second event data indicating the second event based on the correlation and the unknown data when the unknown data is input and output data is output, and for generating the output data indicating the story of the game using the first event data and the second event data;
a play means for causing a user to play the game on a user terminal used by the user who plays the game, based on the output data;
The first event is
a fixed event whose content is set by the first event data and which the user will definitely play,
The second event is
The content of the event is set based on the log data, and the event is variable and may vary depending on the user, and is generated differently for each play.
The second event is an event that occurs between a plurality of the first events.
system.